On Tue, 17 Nov 2015 18:34:06 +0100
Stefan Mark <mark_AT_unserver.de> wrote:
Hey Stefan,
> I know, its all about simplicity, but with only a marginally higher
> complexity an arbitrary number of channels and color spaces could be
> supported. Something like this:
> A type field
> +--------+-----------------------------------------------------------+
> | 1 | Type of channels: 0 predefined, 1 wavelength, 2 names |
> +--------+-----------------------------------------------------------+
> and, depending on the type:
>
> type 0
> +--------+-----------------------------------------------------------+
> | 8 | For example RGB, RGBA, CMYK, LAB, ... |
> +--------+-----------------------------------------------------------+
>
> type 1 (Multispektral Images)
> +--------+-----------------------------------------------------------+
> | 2 | Number of Channels |
> +--------+-----------------------------------------------------------+
> | [8] | List of lower/higher wavelength of channel (2 floats) |
> +--------+-----------------------------------------------------------+
>
> type 2 (Arbitrary image-like data)
> +--------+-----------------------------------------------------------+
> | 2 | Number of Channels |
> +--------+-----------------------------------------------------------+
> | [32] | List of channel labels (32 byte long strings) |
> +--------+-----------------------------------------------------------+
I read your mail at least 10 times now, but I've still got no clue
what the hell you are talking about.
If we go down this path, we probably would end up with the TIFF format
in finite time.
Cheers
FRIGN
--
FRIGN <dev_AT_frign.de>
Received on Tue Nov 17 2015 - 18:43:44 CET